A Variation of the F-Test for Determining Statistical Relevance of Particular Parameters in EXAFS Fits

Abstract

A general problem when fitting EXAFS data is determining whether particular parameters are statistically significant. The F-test is an excellent way of determining relevancy in EXAFS because it only relies on the ratio of the fit residual of two possible models, and therefore the data errors approximately cancel. Although this test is widely used in crystallography (there, it is often called a “Hamilton test”) and has been properly applied to EXAFS data in the past, it is very rarely applied